The finish line is everywhere in this picture
A photo finish records one place over many moments. Build a synthetic finish-line image, change a runner’s speed and discover why strange proportions can still reveal the winner.
The finish line is everywhere
Two runners. Two stretched bodies. One winner. A photo finish ought to settle the argument — until you try to read it like a photograph.
Where is the finish line? Which runner is farther ahead? Why does one body look thinner than the other?
The first question contains the trick. In a line-scan photo finish, every vertical slice records the finish line. Moving across the picture takes you through time.
Build the image below. Watch the runners cross the narrow scanning line, then watch their shapes accumulate in the strip. The strange proportions are your first clue to what this camera keeps — and what it leaves out.
Build a finish, one slice at a time
Use the time control to move slowly through the crossing. The scene shows two simplified runners approaching a fixed line. The accompanying strip collects only what occupies that line at each instant. It has no view of the rest of the track.
In this model, earlier times are on the left and later times are on the right. Follow the labels: an actual published photo finish may run the other way.
Now look for the first appearance of each runner’s torso. Ignore a hand, a foot or a head that arrives earlier. For a running finish under World Athletics rules, the relevant event is the torso reaching the vertical plane at the near edge of the finish line. The other body parts do not decide the order.
The coloured figures here are invented. You are inspecting a controlled final stretch, not replaying a particular race.
Turn the finish line into a clock
Move through the finish, then compare what the camera records. Read the still diagrams for the complete explanation.
Original synthetic model · identical 0.40 m torsos · constant speeds · no real race footage.
A: 8 m/s. B: 10 m/s. Slide or play to collect the finish line.
One moment in space
Now: 0.900 s
One line through time
Only the purple finish line is recorded. Each new column keeps that line at a later instant; nothing beyond the cursor has been collected yet. The dark rectangle is the torso. The hand and head do not decide this running finish.
The recorded figures face left because their fronts cross the line first. In this strip, time runs left to right.
The finish, without moving pictures
Before either torso arrives
The camera watches one vertical line. The runners have equal torso widths but different speeds and starting positions.
1.000 seconds: A reaches the line
A’s front torso edge is on the line. B’s torso is still 0.20 m behind it, even though B is moving faster.
1.020 seconds: B reaches the line
B arrives 0.020 seconds later. In this synthetic finish, A wins.
The full time strip
Read left to right: earlier to later. The first edge of A’s torso is at 1.000 s, B’s at 1.020 s. Compare the torsos, not the hands. A’s 0.40 m torso occupies 0.40 ÷ 8 = 0.050 s. B’s occupies 0.40 ÷ 10 = 0.040 s. Both bodies are the same width in space; B looks narrower in time.
Double the speed, halve the time
With A at 5 m/s and B at 10 m/s, the equal torsos occupy 80 and 40 ms. A’s recorded torso is twice as wide. The arrival times are unchanged: A still wins by 20 ms. We changed A’s initial position to preserve that finish time.
Model limits: rigid two-dimensional figures, constant speeds, fixed poses and a perfectly thin recording line. Real athletes move their limbs and lean; real systems have finite resolution, calibrated timing and officials. This model explains the geometry and does not measure a real result.
The faster runner becomes narrower
The two model torsos have the same width in the direction of travel: 0.40 metres. Runner A moves at 8 metres per second; runner B moves at 10.
A’s torso occupies the scanning line for 0.40 ÷ 8 = 0.050 seconds. B’s occupies it for 0.40 ÷ 10 = 0.040 seconds.
The output gives equal amounts of horizontal space to equal amounts of time. So B’s torso takes up less space in the strip. B has not lost weight. The camera has spent less time recording that body at the line.
Try the comparison with A at 5 metres per second and B still at 10. A now takes 0.080 seconds to pass its torso across the line — twice B’s time. The identical torso becomes twice as wide in the result.
This is why measuring somebody’s apparent body width in a photo finish is a trap. You are measuring the duration of a crossing, scaled into pixels.
Faster at the line does not mean first across it
There is another trap in the demonstration. B moves faster, but A’s torso reaches the line at 1.000 seconds. B reaches it at 1.020. A wins by 0.020 seconds.
Those times are settings of our invented scene, not measured athletic results. The runners begin this short visible stretch at different distances from the line. The model says nothing about their speed or position earlier in a full race.
Move through 1.000 and 1.020 seconds again. The first torso crossing determines the order. How wide each runner subsequently becomes in the strip cannot change that.
Nor can you put a ruler across the finished strip and read the runners’ separation in metres. Its horizontal axis gives elapsed time. A distance calculation would need more information about their motion.
Why a strange-looking image can settle a race
A line-scan camera repeatedly records a very narrow view aligned with the finish. Software places those records beside one another and connects them to a time scale. FinishLynx’s manufacturer describes exactly this assembly of thin images.
A real runner is more complicated than our rigid silhouette. A planted foot, swinging arm and moving torso need not pass the line at the same speed. Different parts can therefore acquire different proportions in the assembled image. The distortion is a consequence of collecting a crossing over time.
The image still needs a correctly aligned camera, a reliable time scale and someone reading the relevant body part. World Athletics’ rules specify checks on the equipment and start signal, and assign officials responsibility for interpreting the finish. An unusual-looking picture is not, by itself, evidence of a bad result.
One question to keep
Our model holds speed and body shape constant during each crossing. It omits stride, leaning, perspective, overlapping runners and camera exposure. Its displayed times explain the construction; the page is not a timing instrument.
Return to the complete strip. Each runner appears in pieces collected at different instants. There never had to be one instant when the whole scene looked like this.
The useful question is small and exact: when did this torso first reach the finish? Once you ask that, the apparently warped picture becomes a clock you can read.